Metraipolis › archive › dc_7e6dd90c2b09
Finite Probes Suffice: Identifiability and Universality for Weight-Space Learning — arXiv:2609.33901
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.33901
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T21:01:53.102Z
- SHA-256 of the archived text
a8a57c8b36afd0a461b0038f9fb8ae71c15476f5eba377bed27e6d801650cccc- Stable URL of this page
https://metraipolis.com/documento/dc_7e6dd90c2b09
How to cite
Finite Probes Suffice: Identifiability and Universality for Weight-Space Learning — arXiv:2609.33901 — arXiv (en), https://arxiv.org/abs/2609.33901 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7e6dd90c2b09 (Metraipolis archive dc_7e6dd90c2b09, sha256 a8a57c8b36afd0a461b0038f9fb8ae71c15476f5eba377bed27e6d801650cccc)
BibTeX
@misc{dc7e6dd90c2b09,
title = {Finite Probes Suffice: Identifiability and Universality for Weight-Space Learning — arXiv:2609.33901},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.33901},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7e6dd90c2b09; sha256 a8a57c8b36afd0a461b0038f9fb8ae71c15476f5eba377bed27e6d801650cccc},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.33901. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Learning properties of neural networks has recently attracted growing interest, with existing approaches operating either directly on network parameters or through probe-based representations of network behavior. While probing methods have shown strong empirical performance, their theoretical foundations remain limited. In this work, we study when finite probe-based representations are sufficient for learning neural functionals. We establish general identification and universality results for probing, and show that using intermediate hidden representations can provide significantly more informative representations than relying only on final outputs. Motivated by these results, we introduce HIDDENPROBE, a simple architecture for learning from hidden probe responses. Across a range of neural functional benchmarks, including both MLPs and Transformers, HIDDENPROBE consistently improves over existing probing methods and achieves state-of-the-art performance. Our code is publicly available on GitHub. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.AI). Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.33901 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-07T20:29:02.779Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).